What I read in 2004

I'm not completely sure whether to call this a book, or a book-shaped art object. It consists of ten very large and detailed cartoons, each the size of a full-sized newspaper page, printed on heavy cardboard, along with an introduction and a "Comic Supplement": the late 19th and early 20th century cartoons that Spiegelman was reading while he drew the cartoons in this book. (Some of these cartoons are available only because of Nicholson Baker's newspaper preservation project.) The cover is a variant of Spiegelman's famous black-on-black New Yorker cover.

These cartoons were originally published by a number of mostly foreign papers, between early 2002 and late 2003, and they're about Spiegelman's own reactions to September 11 and its aftermath, about the mass insanity the attacks brought out in the US and the somewhat different kind of mass insanity that was confined to New York, and, as Spiegelman said in an interview on a local Bay Area radio show, about "the unhinging of one brain in lower Manhatten".

I have the same reaction as most people: this is my least favorite of Austen's novels, and I can't help thinking this is because it's the one I understand the least. It's the one where I'm most aware that I'm reading about an alien society and that I have to make a deliberate effort to see what's motivating these people. I liked it better this time than I have before, though: there still aren't any characters I really admire (again like most modern readers, I just can't warm to Fanny), but there are some characters who are wonderfully drawn as characters. The Crawfords are interesting and flawed, and Mrs. Norris is great satire.

There are a few aspects of society that we see more of in this book than in others: how a naval officer's career might start (something Austen would have known well), what kinds of things people with good literary taste would have been reading, and what it meant for the head of a household to manage his household well or poorly. I think Said and his ilk are overreading this book a little when they talk about the subtext of slavery: yes, Sir Thomas is absent for most of the book to do something unspecified involving his estate in the West Indies, and yes, that estate would have been worked by slaves, but nothing in the book relies on that. The book would have proceeded the same way if he had been absent anywhere far. Still, it's certainly true that Sir Thomas's main role in this book is as an authority figure and that the author emphasizes how important it is for a household to have a strong authority figure.

I understand better than I did why the play was such an improper thing to do (it's not prudery, and it's not a prejudice against the theater), but from a modern perspective it still seems odd. I suppose someday I ought to read "Lover's Vows": some of the objections had to do with the particulars of that play.

The meta-premise of this book: Victorian novels like the ones Jo Walton grew up reading are wonderful, but it's impossible to believe in them. The characters don't act like humans. Especially, the female characters don't act like real women of our species. So she set out to write a book where the axioms of the sentimental Victorian novel have a real biological basis. She wound up with a book that resembles the Barsetshire series more than a little, but that also happens to be a fantasy novel about dragons.

So we have the worthy patriarch, with solid but not exalted rank and adequate but not great wealth, dying in the first chapter, leaving one son established in the Church, one daughter well married to a powerful member of the nobility, one son trying to make his way in the world, and two daughters in need of husbands and protectors. As with Trollope, we visit a number of different social worlds and their interactions with each other: the Church, the country nobility, government, the law. It's a funny premise, but the book is genuinely affecting.

Published posthumously. The cartoons appeared between 1932 and 1989 and were selected by his widow, Tee Addams. This book includes all of my old favorite cartoons and a few I hadn't known before.

I just noticed an inscription in our copy (which we bought used in the East Bay) that makes it extra-cool: "For Decca and Bob — Hope you will like Charlies' Book as much as we liked your American Way of Death. Kindred spirits, we all — Love — Tee."

This book is Another New Crobuzon novel. I wouldn't call it a sequel to Perdido Street Station and The Scar, just as I wouldn't call The Scar a sequel to Perdido Street Station. It takes place (mostly) a few decades after those books, and, while there are a few offhand references to earlier events, they aren't important. There's no real connection other than the city itself.

I found this the strongest of the three New Crobuzion books. It's certainly the most topical, and the most overtly political. It's about war and revolution and labor rights. To me it seemed like an alternate version of 1848, meaning that it's about the cruelty of power, about the awkward struggle to understand that the world can be better tomorrow than it is today, to learn what a just society might look like, about noble dreams and grand gestures, some of which are useless, about betrayal, about the tension between historical forces and individual choices. It's about an imagined world, and about 1848, and about today.

Quantum computing is a new field. The very first paper on it (by Feynman!) was around 20 years ago, and there were a few papers published around the time I started grad school, but it was low enough profile so that, even though this was work by theoretical physicists, I wasn't really conscious of it while I was in grad school. My sense is that most people didn't pay attention to any of this work until Peter Shor's 1994 paper on factorization. The fact that quantum computers give us a polynomial-time algorithm for a problem that is widely believed (not proven, just widely believed) to have no classical polynomial-time solution is enough to make this look interesting!

There still aren't any practical applications of quantum computing. There's been just enough work on the experimental side so it no longer seems ridiculous to imagine that someone might someday build a genuine quantum computer, but it's still true that nobody has built one. (Sorry, a special-purpose system that "factors" 15 into 5 × 3 doesn't count.) But even though nobody has built a quantum computer yet, we're now talking about a field with a huge body of research literature both from the physics and the computer science side. It's now a mature enough field to have more than one standard introductory textbook. This is one of them.

I would describe this book as a fairly dense graduate-level introduction. It's less than 200 pages long, and that includes a long appendix reviewing the foundations of nonrelativistic quantum mechanics and another long appendix reviewing a number of different fields of mathematics. (Group theory, number theory, linear algebra, and so on.) It reviews the theory of computation in terms of classical Turing machines, builds up a couple different theoretical models of quantum computation, and illustrates a number of quantum algorithms—including Shor's algorithm, naturally.

Quantum computation is a funny thing. What all of the algorithms have in common is that they involve setting up some initial superposition, applying one or more unitary operators, doing a measurement, and then doing some more classical computation. What's funny about it is that the quantum part of the computation is essentially analog: we usually describe the quantum state as a superposition Σ c_i |X_i>, where the basis states are things like |0010>, and the computation is on the coefficients c_i. The whole reason this can be interesting is that when you perform a unitary transformation on an N-qbit state (i.e. a state of N entangled spin-½ particles) you're dealing with 2^N complex coefficients. But, of course, the really odd part is that even though the computation is on the coefficients, they're unobservable: this is still ordinary quantum mechanics, so the only way you get to see anything is to make a measurement, and the only thing you ever see is some eigenvalue of the operator corresponding to whatever observable you've chosen.

Most of the quantum algorithms turn out to involve clever transformations that boost the amplitudes of the states you're interested in, so there's a high probability that when you make an observation you'll find the state you wanted. Shor's algorithm is a fairly straightforward application of a quantum version of the discrete fourier transform—straightforward, at least, if you happen to know enough number theory and group theory to know that factorization can be reduced to the problem of finding the periodicity of a particular function.

The part I found most unsatisfactory was the theoretical model of quantum computation. The algorithms themselves involve applying one unitary transformation after another, and yes, this can be shown to involve circuits that one can imagine might be physically realizable, but that still doesn't seem to me like what I think of as a model of computation. The book does establish the notion of a quantum Turing machine but it doesn't really use that notion, largely for the perfectly good reason that the physical requirement of unitarity (which of course implies reversibility) makes the quantum Turing machine very inconvenient theoretically. We need some kind of better model for quantum computation, one that's more useful than the quantum Turing machine and less ad hoc than quantum circuits. Perhaps if I keep reading the literature I'll find that someone has come up with such a model.

One of my favorites of her books. Persuasion is about second chances, the possibility of revisiting a seemingly irrevocable decision. It's also the book where the war is most prominent; in the other books you'll only notice it if you know to look. But here it's very easy to see that you're reading a book written soon after a time of national crisis, by an author with close relatives in the Navy.

A minor detail that I noticed this time; I can't remember whether I'd noticed it before. In the last few chapters there are a few occasions where Anne decides, for trivial reasons, to postpone telling anyone what she has learned about Mr. Elliot. Now all of these mentions are brief, and there's nothing to call them out as important. But they seem like the sort of thing where an author is deliberately not calling an action out as important in order to build suspense, to show that a seemingly trivial decision has momentous consequences later. It's reminiscent of Elizabeth Bennet's decision not to tell anyone about Wickham. But the difference is that, as far as I can tell, Anne's delay in telling anyone about Mr. Elliot really is as trivial as she thought it would be. There are no consequences. Lady Russell would have done nothing different if she had known. So why the hints? Just misdirection, or a remnant of some other direction that Austen was thinking of in an earlier version?

After the election, I figured I needed a Discworld book. This one is mostly about elves, who, as is often the case in the authentic fairy tales, are decidedly Not Nice.

Reviews of the Discworld books usually point out that one of the reasons they're worth reading is that they discuss serious issues using the techniques of humor and fantasy. But it occurred to me, reading this book, that one of main the reasons I like them, one of the reasons the humor works, is that they're textually dense. They're like the classic Warner Brothers cartoons: lots of the jokes require you to know something specific, but you might not notice them if you didn't happen to know it. It's a perfect example of "the more you know, the more jokes you'll get." I enjoyed spotting the allusions to A Midsummer Night's Dream, Henry V, J. R. R. Tolkien, Lord Dunsany, "Tam Lin", Dirty Harry, P. G. Wodehouse, the sayings of Isaac Newton, and probably others that I've already forgotten. (Not counting the things I didn't catch!)

One of Stephen Jay Gould's last books, published posthumously. It's a full-length book, not a collection of essays. The title refers to the ancient proverb that most of us probably now know through Isaiah Berlin: "The fox knows many things, but the hedgehog knows one big thing." The subtitle of this book is "Mending the Gap between Science and the Humanities," and, indeed, that's what the book is about: the proper relationship between sciences and humanities. Gould partly demonstrates this through example, discussing scientific topics by close textual analysis of original sources, focusing especially on incunabulae from his own private collection. (He must have come to that interest late in life; I've read most of his earlier books, and I'm sure that if he had been a serious book collector earlier, he would have mentioned it.) Some of this is fascinating. I was delighted to learn what the trite phrase "sweetness and light" actually referred to!

Unfortunately, I found the central argument of this book seriously lacking. There were two main reasons for that.

First, Gould never quite says what he means by "science" and "humanities". This might seem trivial, but it really isn't. I think I have some idea what he means by science (the two fields he discusses most are particle physics and paleontology, and that's enough to give a pretty good idea of what he's talking about), but "humanities" is a lot harder to grasp. Many times in the book, Gould refers to science and humanities as two major currents of thought. (Which, incidentally, he does not identify with the fox and the hedgehog; he uses that metaphor for something else.) But two and only two? I just don't see that much in common between an economic historian, a classicist, an avant-guard composer, an analytic philosopher, a literary novelist, a follower of Derrida in an English department, an anthropologist, and a film scholar. Gould talks about "humanities" as if it's just one thing, but I don't see that all of those people have more in common with each other than they do with your average geologist or astronomer. Early in the book Gould warns about the trap of looking at the world in terms of dichotomies, he warns that some of our terms reify arbitrary and meaningless categories, he warns that the way 21st century universities classify types of knowledge is contingent and political, but then he falls into that trap himself. That's a shame.

Second and more seriously, although Gould talks about "mending the gap" between science and humanities, he never makes it clear what problem he thinks he's solving. In one early chapter he discusses several supposed examples of science versus humanities conflicts: the 17th century conflict between the ancients and the new learning (which, as Gould points out, was a real intellectual conflict but not science versus humanities; it would be incorrect to identify either side in that dispute with either science or humanities), C. P. Snow's famous "two cultures" (which Gould correctly dismisses as overly simplistic while making the same errors himself), and the modern "science wars" (which Gould correctly says aren't as important as some people think, although I'm not sure he correctly identified the participants; he discusses historians and sociologists of science but I didn't see terms like "post-structuralist" or "deconstructionism", or names like Derrida, Spivak, or Feyerabend. He says nobody but marginal figures take a radical relativist stance but he doesn't say whether that means he thinks people like the ones I mention aren't radical relativists, or whether that means he thinks they're marginal. Either position might be defensible, but either would need some real discussion.) But then comes the question: if all of these examples of science/humanities conflict are mythical or overblown, then what is the gap that Gould thinks is in need of mending? In what way would intellectual life in a better world differ from intellectual life of today?

Any new book by Gould is a treat, and we're not going to get many more. I found this book frustrating, but I still enjoyed it for the many small insights and for the tour into corners of intellectual history that I wouldn't have known about otherwise.

The Book of Honor for Potlatch 14!

I'm not a Dick completist; I've probably read about a half dozen of his books, so I don't have all that much to compare this to. The book I was reminded of most from reading this was Naked Lunch, mostly because both books are unflinching pictures of drug culture that never lose sympathy with the people who live in it. On the other hand, the differences between the two books are equally obvious. A Scanner Darkly isn't about the lust for dominance or about paranoia; it's just about people who are trying to live in a world that doesn't make sense. It's no less sympathetic toward the narks than toward the junkies. There are hints of some vast conspiracy, but it's never really the point. The point is people's individual choices and losses. A few places in the book hint at a possible transcendence, but we don't quite live in a world where such things are accessible.

One puzzle I've been wondering about: the book was published in the 70s, and it's set in '94. There are a few futuristic touches, but only a few. Is this book science fiction? If so, in what sense is it SF and why does it have to be?

We saw the San Jose Rep production last week, so I figured I had to reread this. It's one of my favorite Shaw plays, and with Shaw there's so much that you don't get if you only see the play and don't read it! The prefaces, if nothing else. Shaw describes Andrew Undershaft as the hero of the play, and that doesn't come as a surprise: Undershaft has all the best lines and wins all the arguments. So then comes the question: what's a socialist doing writing a play where the hero is a millionaire who says that the most important moral activity is to acquire money and gunpowder? And the answer, of course, is that Shaw enjoys paradox, he enjoys stating opnions in a shocking way that only gets explained later, and he does explain this later. The explanation is explicit in the preface, but you can see it even in the play itself.

Incidentally, am I the only person who thinks that the arguments Undershaft used to convince Adolphus were really weak? I suppose Adolphus made the right decision, but he shouldn't have made it for the reasons he was given.

An introduction to functional programming with Caml, a dialect of ML, with applications to nontrivial problem domains: regular expressions, parsing, geometric computations, evaluation and type analysis, and so on. It ended up presenting all the pieces you'd need to build a real ML implementation.

I was a little disappointed that this book didn't show examples of how to use ML's module system to build up large programs, though. These case studies are poor examples of building up higher-level pieces from lower-level ones: layers of abstraction weren't cleanly separated, and in many cases the authors had to rewrite low-level components they'd just presented because the needs of the client that used them changed. I don't think this is an indictment of ML; I think it's just that the authors deliberately chose to concentrate on the functional features of the language, not the program structuring features. Still disappointing.

A graduate level introduction to type theory. This book starts at the beginning (operational semantics, the pure λ-calculus) and uses that as a framework to discuss increasingly sophisticated type systems. We end up getting through several versions of parametric polymorphism, several versions of subtyping, the combination of subtyping and parametric polymorphism, existential types, and so on. Starting with the λ-calculus meant that everything is either viewed as syntactic sugar for something that's already there in the λ-calculus or as a simple extension, or sometimes as both. The book is both rigorous and practical: rigorous, in that it's all about proving theorems about type systems, and practical in that all of the type systems come with concrete implementations and in that the book discusses how to implement all of the type checking algorithms.

So what have I learned from this book? Well, I think what I've mostly learned is how type theorists think. I've learned how to give a formal definition of a simple type system, how to prove theorems about a type system, and, most importantly, what kind of theorems you want to prove: progress, preservation, minimal typing, and so on. I've got some idea of what properties you want a type system to have. I've been exposed to a few new idea about types, most notably kinds, a.k.a System F_ω, and existential types. (My first thought was that kinds were reminiscent of C++ class templates or ML functors, but it hadn't occurred to me to think of kinds as an infinite hierarchy.) I've learned to think more carefully about some distinctions that I'd only vaguely thought about before: nominal versus structural typing, finite versus recursive types, and so on.

Now that I've read this book, I should have no trouble reading the type theory research literature.

A number of Austen characters are mistaken about important things, but one of the interesting facts about Emma is that Emma turns out to be mistaken about practically everything. Obvious but unspoken social nuances turn out to be less obvious than she thinks. This book shows how an intelligent and perceptive person, due partly to wishful thinking, partly to excessive certainty and refusal to reexamine prior judgments, and partly to being just slightly less perceptive than she thinks, can continually delude herself. It's interesting that most of the other characters in the book misestimate Emma's judgment the same way she does, and convince themselves that she must somehow know things they haven't told her.

It occurred to me that only a small change of perspective would be required to change the name of this book from Emma to Jane Fairfax.

A group of friends in the Davis area get together to talk about Jane Austen's novels, with some (loose) parallels between the events in their lives and the events in Austen's books. The characters seemed real; they reminded me of the people I know who like to talk about Austen and books in general, and they admire many of the same authors that my friends do. The book is written in the first person plural; I can't recall reading another book that did that.

This book got a lot of attention when it came out, which it deserves. Make sure not to skip the ancillary material at the end! (Brief summaries of Austen's novels, quotes from people about them, and "discussion questions" posed by each of the characters.)

Now I'll have to go reread Austen.

We probably aren't going to do a domestic open adoption, but this is the best book I've seen about the subject. It covers many aspects of open adoption, and it's the first book I've seen that gave me a good sense of the history. (It's the first book I've seen, for example, that acknowledge that open adoption is now the default for domestic non-stepparent adoption and that the sorts of adoptions that were common in the first half of the 20th century are now quite rare, and that explained how that change happened.)

Most of the material in this book ends up being about managing the relationships between the adoptive family and the birth parents. This may not be an analogy that would have occurred to the book's authors, but it reminded me of the sorts of things I hear from my friends in the poly community.

Partly informative, partly annoying. The author is a pragmatist in the tradition of Dewey. He sees that tradition as related in some ways to logical positivism (in that it "answers" some traditional questions of philosophy by saying that they aren't worth asking) and related in other ways to the tradition that he describes as "textualism": people like Derrida and Foucault. The book is a collection of previously published essays, with a long introduction on what pragmatism is.

In the introduction, and elsewhere, I found myself having a curious double reaction: partly "well, of course this is true; it's trivial!" and partly "how can anyone expect me to swallow an outrageous assertion like that?". Curous because I often had both reactions in response to the same assertion! One reason is probably just because this isn't the kind of philosophy that I'm most sympathetic to. But another, I think, is that Rorty indulges in a semantic game where he says something that can be interpreted multiple ways, one tame and trivial, another far-reaching and dubious.

The parts of this book I found most interesting were the essentially sociological ones: the history of American university philosophy departments over the last 60 or 70 years, what contemporary philosophy professors do and how they see themselves, how academic philosophy today relates to other disciplines. (Physics; literary criticism.) The part I found the most annoying was his discussion of Derrida. I think Rorty was trying to show, by example, the style of discourse that he suggested might be more common among post-philosophers who accept pragmatism than it is among philosophers today. What is one to make of "The Freudian distinction between the normal and the abnormal, drawn with the concreteness which is given by Derrida's exhibition of the sexual overtones of most metaphilosophical debate, seems to be just what is needed to be properly playful about the difference between the Kantians and the dialecticians."? One might point out that Rorty doesn't even attempt to show that Freud's ideas about sex have any validity, or that Derrida's "sexual overtones" are really there. But then, that would be missing the point. Ideas like "validity" and "really there" are just what's being rejected here; this passage is being written about, and from, a perspective where choosing a new vocabulary is supposed to be more interesting than constructing an argument for an idea. But even with that said, what is one to make of a passage like this? If you don't already accept the presuppositions of that passage, are you just supposed to be dazzled by the rapidity of the passage from one topic by another, or by what Rorty calls "name dropping"? If so, then about the only thing I can think of to say is that it's a matter of taste: either you find that sort of thing dazzling or you don't. I don't.

But that may be a little harsh. After reading this book I have a slightly better grasp than I did before of what it means for there to be philosophical traditions that just ignore issues that I think of as central, that don't think it's interesting to formulate an idea and formulate an argument in support of that idea, that have decided to play some other game entirely. It's still not my idea of what's interesting, but I understand a bit more why some people might think it is.

I've finished Remembrance of Things Past! (And now we've even had the celebratory dinner party.)

Time Regained takes us through the war and well beyond. It's set in two distinct times: once when Marcel returns to Paris during the war (in 1918, I think) for a relatively brief visit after some years in a sanatorium, and then when he returns again some time in the mid-to-late 20s after more years in the sanatorium. Each time, and particularly the latter, he notices the changes: in customs, in the arrangements of the social world, in his friends. Some have grown, some have grown old, some have died. Is this the collapse of an old world or is it just the ordinary process of change and renewal, which seems more momentous than it is because of the illusion that a world has always been the way it happens to be at the moment one first sees it?

Most of the second half of the book takes place at a party given by the Prince de Guermantes. (Whose new wife is the former Mme. Verdurin!) The narrator's first thought is that everyone he knew was in disguise; then he realizes that they're old; then he realizes that he is old too. So part of what this book is about is the passage of time, the situation of people in time as well as space.

This book also takes us back to the beginning, showing us the connections between all the different threads of the entire work, how they grew out of the events at the very beginning, including the two ways at Combray, Swann's Way and the Guermantes Way. This is one of those works where the book itself tells you how to read it, with the weird twist that it doesn't tell you that until the last hundred pages of a work that's more than three thousand pages long.

There are, I think, two crucial sections of this book that say what the narrator's view of literature is and what kind of book this is supposed to be. First, when he reads a passage from the Journal of the Goncourts while visiting his childhood friend Gilberte de Saint-Loup (née Swann). Second, an epiphany he has while waiting to go in to the Prince de Guermantes's party. In the first of those scenes he is depressed because he thinks that passage shows him that he has no talent for literature and that even if he did it wouldn't matter, because that passage also shows him that literature is false and trivial. After the epiphany, however, which is reminiscent of that infamous madeleine, he realizes that literature can be something else, something that is worth doing. I won't try to summarize what he has come to realize literature can and should be, but one very crude way of expressing something like it would be to disginguish between realism and minute observation of the world versus subjectivity and observation of the internal.

Someone, I've forgotten who, made the claim that Remembrance of Things Past actually has a simple plot: the uncertainty about whether Marcel will ever become a writer. At the end of this book, he does.

My friend Alex Stepanov has been telling me to read this book for literally years. I finally gave up on buying a copy (copies of the English translation are scarce), and found one at the UCSF library. I'm the first person to have checked it out in thirty years.

Bourbaki, of course, was the most famous nonexistent French mathematician of the last century, and Elements of Mathematics is an attempt to construct a rigorous and self-contained description of mathematics starting from a purely lexical formulation of logic and progressing from there to set theory and beyond. (To get to an idea of how formal this is: it isn't until page 173 that we get to the theorem If a and b are integers, then a < b iff ∃ c > 0 such that b = a + c.)

Actually, Alex just told me to read chapter 4 (structures). But I'm stubborn, so I also read the chapters on logic, set theory, and the construction of cardinals and integers from set theory. The thing I found most striking in those chapters was the use of the τ operator as one of the primitives of logic. τ does double duty: it can be used to define "∃" and "∀" (I'm used to thinking of one of them as a logical primitive), and, when imported into set theory, it gives you the same things you'd get from the Axiom of Choice. I'm not so sure I like that: I don't have anything against the Axiom of Choice, but it seems odd to build mathematics using a version of logic that makes it impossible to define set theory without it.

I can see why Alex thinks that chapter 4 is relevant to generic programming: the notion of a "species of structure" (roughly: a typification built up from a base set, combined with an axiom) seems very much like what people who do generic programming call a "concept". I'm not sure if there's anything in generic programming that's directly equivalent to morphisms, which are the other major subject of chapter 4.

Incidentally, a minor mystery: I can't find anything in this book to say who translated it. I suppose my best guess is that Bourbaki translated it themselves.

The literal referent of the title is, again, Albertine, who left Marcel at the end of The Prisoner. He spends the first part of the book scheming to get her back. This soon becomes impossible when Albertine dies in an accident, so most of the book is about his coming to terms with her death.

Part of this is his continued obsession about whether Albertine was lesbian. He finds conclusive proof that, yes, she was sleeping with other women during much of their affair. But then he finds good reasons to wonder whether there might be other explanations, however unlikely, and whether his conclusive proof is really conclusive after all, and finally comes back to the impossibility of fully understanding another person or an event of the past. (We see something similar, later in the book, when he tries to understand the significance, in light of later knowledge, of small scenes with Saint-Loup.)

He also, I think, comes to some acceptance of his own bad behavior. If Albertine concealed the fact that she was lesbian (or more likely bisexual, although I'm not sure that the word or the concept was around in Proust's time), it was largely because of his own jealousy, suspicion, and tyranny. So much of his behavior toward her was based on lies and petty dominance games! But after her death he is genuinely grieving. Compared to the relationship itself, it's quite sweet.

Finally, this book is about the process of forgetting: the stages by which the narrator stops grieving for Albertine and becomes ready to move on. (Or at least as ready as he ever has been or will be, which by most people's standards isn't very.)

Part of a series of detective novels set in Tokugawa-era Japan. I've enjoyed previous books set in that series. (Digression: a friend once asked, when I told him what one of those books was about: why would I want to read a book where the main character is working for an unpleasant police state? The answer, of course, is that that's the point! It's not at all uncommon for a detective story to be about the trials of being the least corrupt person in a thoroughly corrupt world. You get that in hard boiled detective novels, for example.)

Unfortunately, I think it's time for me to stop reading this series. In this book the author succumbed to temptation in several different ways: too much of a 20th century sensibility, too much emphasis on events that would have had major historical significance, and way too much magic powers.

Jacobs describes a dark age as cultural amnesia and extinction: a culture is destroyed because the knowledge that's required to systain it is lost. The dark age everyone thinks of first, and the one Jacobs talks about the most, is the collapse of the Western Roman Empire. I have some quibbles about the way Jacobs interprets the history of late antiquity and the early Middle Ages, but none of that is really important to her arguments. Her real concern isn't the late Roman Empire, but early 21st century North America.

Jane Jacobs is most famous for The Death and Life of Great American Cities, and it's still her best book. It's a close study of how urban communities work, what the preconditions for a successful community are, what a successful community looks like. Dark Age Ahead, more than any of Jacobs's intervening books, reads as a return to The Death and Life of Great American Cities. She's looking at many of the same as in the earlier book and asking: what has changed in the intervening forty years? And the answer is that most of us failed to learn from her earlier book. The dark age that Jacobs writes about, the loss of knowledge, is the loss of knowledge of how to live in local communities. As she points out: someone who lives in a functioning community that's under threat will know to defend it, but someone who has never had that experience will never know what they're missing.

Not really a good book (the characters are flat and appear out of nowhere; the mystery involves that stalest of devices, an insane criminal), but a curiosity because of the time and place where it's set: British East Africa, in the late 30s. Some of the characters are German expatriates, and part of the plot revolves around internal power structures in the African branches of Nazi front groups. At the time the book was written there was tension, but not war, between England and Germany. A moment earlier or later and it couldn't have been written.

And yes, the author was related to those Huxleys.

Similar to his marvelous short story, "A Colder War". It's not exactly set in the same world, but the spirit is the same: the intersection between bureaucracy, spy thriller, technology, and Lovecraftian horror. It's an original and striking combination, and the horrors become more horrible when they're seen in this context. There's a reason for that phrase "banality of evil".

Oh, and this book also includes an explanation for why volume 4 of Knuth hasn't been published yet.

One of the basic facts about American politics today is that most white working-class men vote against their own economic self-interest. This book attempts to explain that, and to explain why American politics deals so much with symbolic issues rather than basic matters of money and power. This book is an example of good reporting: illuminating a broad and important question by close observation of specific people in a specific time and place.

I also recommend the journal Thomas Frank founded, The Baffler.

(I realized, by the way, reading this book: Frank and I were in high school at exactly the same time, and he and I were both high school debaters. Kansas and Pennsylvania aren't all that close, but they also aren't all that far. I wonder if he and I might have once fought over effluent fees for water pollution, or self-extinguishing cigarettes, or tuition tax credits, or arms sales to Morocco?)

Winner of the 2003 Tiptree Award!

I don't want to say too much about this book, because I don't want to give too much away, except: read it. It's a really good novel. Well, I guess I can say one thing. This book wasn't published as science fiction, but it does something that's central to SF even though very few SF writers have managed to pull it off. Campbell supposedly challenged writers to "Write me a story about an organism that thinks as well as a man but not like a man." That's a tough challenge, but Matt Ruff did it. His book (and here I'm not giving anything away: this is revealed on the jacket cover and in the very first page of the book) is about someone with multiple personalities. The main character is mostly functional, but he does not have the same sort of mind that I do. This is one of the most convincing portraits of an alien intelligence I've ever seen.

Is this an accurate picture of how someone with multiple personality disorder thinks, of the causes and management of the condition? No clue. (And apparently there's no consensus about any of those things in the psychological community.) Regardless of that, it's a compelling story.

This book has made headlines largely because of the parts of chapter 11 that describe the Bush administration's single-minded focus on Iraq, and its disastrous decision to invade. Many press reports have quoted the parts where Clarke describes the administration's foolish insistance that Iraq must somehow have been involved in September 11 and the parts where Clarke has explained just why the Iraq war was such a bad idea, the ways in which it helped al Qaeda and hurt our national security.

All of that is true, and important. But this isn't even all of chapter 11, let alone the book as a whole. Chapter 11 is a devastating portrait of an Administration that had no serious response to a national security threat, that played political games instead of solving problems. We have made no substantive improvements in domestic security, and many of the administration's highest profile actions (the poorly named "Patriot" act; the Homeland Security department; the illegal detentions of US citizens) have made things even worse than they were before September 11.

But again, that's only one chapter. The rest of the book is a careful analysis of how the US government slowly learned about al Qaeda, how various administrations tried to respond to it. It's a fascinating portrait of how Washington bureaucracies work, how agencies can promote or thwart policies.

This is a book on an important subject by someone who is honest, serious, and knowledgeable. Everyone should read it.

A standalone book, but one that deals with many of the same concerns as Ken MacLeod's two earlier series. (Fall Revolution and Engines of Light.) Most obviously, it deals with the Singularity and posthumanism. The book, like parts of the Fall Revolution series, is set in a post-Singularity world, and, as in the other books, the viewpoint characters are humans, not posthumans. The humans believe that a Singularity is a dead end, that it is collapse and chaos, not transcendence. One of the terms for it in this book is the Hard Rapture: a deliberate echo, I'm sure, of The Star Fraction's dismissal of the yearning for Singularity and transcencence as "the rapture of the nerds". But then, it's not clear that we should believe the humans! Or, if we do believe the humans, it's not clear which ones to believe: there are at least five human cultures in this book, all of whom have different values and goals. Which of course is another concern that's common in Ken's fiction: the wide spectrum of ways in which human societies can be organized.

One of the chief virtues of good science fiction is that it can expand our sense of the possible, and one of the chief virtues of Ken MacLeod's fiction is the way he applies that to political, economic, and social ideologies: his fiction usually treats all sides of an ideological conflict with respect. Tom Shippey suggested an exercise when you read the Silmarillion: after each disaster you should try to trace through the answer to the question: whose fault was it? Similarly, when you read this book (or any of Ken's other books), a good question to ask every so often is: which side should I be on? You usually can't come up with a definite answer, but it's important to ask the question; you can't just assume that the main character is on the right side. Sort of like real life, perhaps.

Thomas Friedman has a reputation for thoughtfulness—a reputation, in my opinion, that is almost completely undeserved. I read this book because I thought it was on an important subject and that I should read thoughtful people I disagreed with. That's not what I found. This book is a collection of columns. I read some columns I disagreed with, some columns I agreed with, some that were just plain silly. I read no columns that exposed me to a way of thinking that I hadn't seen before, or that showed me that a particular issue was more complicated than I had realized. All of these columns were shallow. Friedman has no capacity for analysis, no capacity for writing about nuances, and, worst of all, no capacity for avoiding intellectual ruts. He proudly writes about traveling thousands of miles and talking to exotic natives, but somehow all he hears from them is that he's always been right about everything all along.

Part of what makes this such a thoroughly bad book is the column form itself: when each chunk is just a page and a half long, it's hard to go beyond familiar ideas and cliches. Try explaining something in a page and a half that readers haven't already heard! It can be done, but it takes real skill from the writer. And it probably can't be done if you indulge yourself the way Friedman does: the name dropping to make sure we know he gets to talk to important people, the quotes from admiring foreigners with slightly broken English, the cutesy phrases he loves to make up. ("America Online" versus "America Onduty"; "God 2.01"; "Naked Air".)

I was going to say I didn't read a single column that taught me anything, but that's not quite true. I read exactly one column that I learned something from: "The End of NATO", which was inspired by a flight out of Afghanistan and had some interesting points about the qualitative technological gap between the US military and the other NATO militaries. It made me think that Friedman could probably be a competent reporter if he decided to make an effort: if he decided to can his rhetorical tics, and to observe and listen rather than to treat interviews as a way of demonstrating his own importance and cleverness.

The actual subject of this book is more specific than the title would suggest. It has a little bit about "globalization" in general, but essentially it's a sustained attack on the IMF's policy toward developing nations. It argues that the IMF is dogmatic, imperious, and undemocratic, that the IMF forces developing nations to implement economic and political policies that no nation, including the US, would choose for itself (e.g. raising interest rates and cutting spending during a recession), that its policies are unchanging doctrine rather than analysis based on the specific history and conditions of each individual nation, and that this doctrine is based on economic ideas that have been discredited for decades. The book suggests that the IMF behaves this way partly because of an unquestioned ideology of free-market fundamentalism and partly because, for structural and historical reasons, the IMF represents large countries' financial industries more than it represents those countries, or the world, as a whole.

For me, this book answered some questions and raised others. I have a better understanding than I did of what the IMF is, what its charter says it should do, and what it actually is doing. I still don't understand why the IMF is as powerful as this. Why does the IMF still have so much credibility that its actions determine the actions of other lenders? Why don't the World Bank, and private lenders, just ignore the IMF as a bunch of incompetent ideologues and cranks?

More Proust! As usual, the title of this book is both literal and metaphorical. This is the book where the main character enters aristocratic society, of which the Guermantes family is an exemplar. Part of the point is that "society" is an illusion, or at least a matter of perspective: we see the same people from multiple point of view, and we see that different people have alternative ideas about who matters, what constitutes social success, what kinds of power and fame are real.

Two interesting questions, which I think I partly know the answers to but only partly: what does the narrator see in fashionable society, and what does it see in him? A lot of this book works by indirection!

John Mitchell teaches at Stanford, and I suspect this book started out as notes for an undergrad class. I liked it a lot: a good overview of the underlying theory behind most high-level languages, and of the issues and tradeoffs that go into designing a language. He covers some theoretical issues (λ-calculus, type theory, formal analysis of concurrency), but it's much more a practical book than a theoretical one. Mostly, what I imagine you'll get out of it is a good idea of the diversity of different languages.

The book starts out with functional languages like ML and lisp (not surprising), and goes on to imperative languages like Algol, languages that support data abstraction, and object-oriented languages. Heavy coverage of object-oriented languages: it made up four chapters, and used four languages as detailed examples (Simula, C++, Java, and Smalltalk). I slightly regret that emphasis: object-oriented languages are popular nowadays, but they're hardly the only model out there. Still, I can understand doing things that way. There are a lot of different ideas about what "object-oriented" means, and lots of different kinds of implementations. It's interesting to see how the different design goals for C++ and Smalltalk affected the way the languages were designed and implemented.

The book finally closes with a chapter on concurrency and one on logic programming. Those are the chapters that I wish had been more detailed, because they were the least familiar to me. I'm familiar with the C/C++/Java version of concurrency (threads, mutexes, semaphores, etc.). This book presented some alternative models, including the Actor model and the version of functional concurrency developed for Concurrent ML, but both were sketchy enough that I had trouble understanding the details. Ditto for logic programming: I think I need more than a chapter to understand what kinds of problems logic programming is useful for and why Prolog was designed the way it was.

Still: if you're looking for an introductory book on language design, I'd recommend this one. Unlike a lot of other language design books I've seen, this one makes a real effort to cover the spectrum of languages that exist today.